agentsclimarketplace

4plus1 models

Skill MarieLynneBlock/arcanum-artifex/workflows/4plus1-diagrams/skills/4plus1-models

Prompts, skills, and agents that survive contact with real workflows. No vendor loyalty. Occasionally heretical. πŸ§™πŸ»β€β™€οΈ

Install
npx -y skills add MarieLynneBlock/arcanum-artifex --skill 4plus1-models

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Produce Philippe Kruchten's 4+1 architectural view model for a software system. This is the core method skill: audience routing, concerns, per-view generation, and cross-view consistency. It outputs canonical diagram-as-code (Mermaid / PlantUML) plus view prose, and does not own draw.io or Miro rendering.

SKILL.md

5.4 KB, as published. Nobody here has run it

4+1 Models (Core)

Generates Kruchten's 4+1 View Model for any software system.

This skill owns method logic only:

  • invocation mode and audience routing
  • context and concern capture
  • per-view architecture content
  • canonical diagram-as-code (Mermaid / PlantUML)
  • cross-view consistency

It does not own output-track rendering mechanics for draw.io or Miro.

The five views at a glance

ViewWho reads itWhat it answersPrimary notation
LogicalEnd-users, analysts, architectsWhat components exist and how they relateMermaid class / component / C4 Container
ProcessIntegrators, performance engineers, operations, BAsHow the system behaves at runtimeMermaid sequenceDiagram (dev) / flowchart with swimlanes (cross-functional)
DevelopmentDevelopers, software managersHow the codebase is organisedMermaid flowchart / C4 Component
PhysicalSRE, infrastructure engineersHow it's deployed and operatedPlantUML deployment + AWS/Azure/GCP stdlib (primary) / Mermaid C4Deployment (fallback)
Scenarios (+1)All stakeholdersKey use cases that exercise the other fourMermaid flowchart (use-case style) + mini sequences

Reference files per view: references/logical-view.md, references/process-view.md, references/development-view.md, references/physical-view.md, references/scenarios-view.md.

Workflow

Step 1 β€” Determine invocation mode

Decide which mode the user is in. If ambiguous, ask.

  • Zero-input mode β€” user gives a short description and expects a draft with explicit assumptions.
  • Interview mode β€” user has real context and wants rigour.
  • Partial mode β€” user wants only one or a subset of views.

Step 2 β€” Determine audience (ALWAYS ASK unless user pre-stated it)

Ask if missing:

"Who's the primary audience for this documentation? (a) Dev-only (b) Cross-functional (c) Executive"

Audience determines notation choice per view.

Step 3 β€” Gather system context

Gather context with one concise checklist first, then only targeted follow-ups if needed.

Start with at most these fields:

  • System name and purpose
  • Stakeholders and concerns
  • Tech stack
  • Scale profile
  • Quality attributes
  • Constraints
  • Out of scope

Mark assumptions explicitly using > **Assumption:**.

Questioning rules:

  • Keep the first interaction to one grouped question block.
  • If the user gives partial info, proceed with assumptions rather than asking every missing detail immediately.
  • Ask one follow-up at a time only when a missing answer blocks the next deliverable.
  • Use plain wording and provide concise options where possible.

Step 4 β€” Route concerns into each view

Use concerns/README.md to select relevant concern modules.

Default concerns:

  • gdpr-data-protection.md
  • security.md
  • bias-fairness.md (if ML present)
  • regulatory-compliance.md (if regulated domain)

Step 5 β€” Generate each view

Generate in order:

  1. Logical
  2. Process
  3. Development
  4. Physical
  5. Scenarios (+1)

Use templates/view-template.md.

Step 6 β€” Cross-view consistency check

Verify:

  • consistent naming across views
  • process behaviours map to deployable units
  • scenarios exercise elements from core views

If writing to disk, run:

python scripts/validate-views.py <output-directory>

For example:

python scripts/validate-views.py docs/architecture

Step 7 β€” Output format (core)

[system-name]-architecture/
β”œβ”€β”€ 00-system-context.md
β”œβ”€β”€ 01-logical-view.md
β”œβ”€β”€ 02-process-view.md
β”œβ”€β”€ 03-development-view.md
β”œβ”€β”€ 04-physical-view.md
β”œβ”€β”€ 05-scenarios-view.md
└── diagrams/
    └── mermaid/
        β”œβ”€β”€ logical-view.mmd
        β”œβ”€β”€ process-view.mmd
        β”œβ”€β”€ development-view.mmd
        β”œβ”€β”€ physical-view.puml
        └── scenarios-view.mmd

When used inside the 4plus1-diagrams workflow, the visual-format skill adds a diagrams/drawio/ or diagrams/miro/ sibling folder alongside diagrams/mermaid/.

Quality standards

  • No placeholders
  • Diagram + prose for every view
  • Rationale tied to quality attributes/constraints
  • Audience statement at top of each view
  • Visible assumptions
  • Specific concerns only

Reference index

Per-view detail:

  • references/logical-view.md
  • references/process-view.md
  • references/development-view.md
  • references/physical-view.md
  • references/scenarios-view.md

Notation cheatsheets:

  • references/notation-mermaid.md
  • references/notation-plantuml.md
  • references/notation-bpmn-in-mermaid.md

Concerns:

  • concerns/README.md
  • concerns/gdpr-data-protection.md
  • concerns/security.md
  • concerns/bias-fairness.md
  • concerns/regulatory-compliance.md
  • concerns/sustainability-climate.md
  • concerns/accessibility.md

Template:

  • templates/view-template.md

Worked example (core views):

  • examples/synth-claim/

Scripts:

  • scripts/validate-views.py

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.